CHIKV 28dpi scRNA-seq
Ryan Sheridan
2024-01-23
Summary
Macrophages
- Lining macrophage clusters were further assessed. The detection of this population is inconsistent for the mock-infected replicates (higher for r1, really low for r2, r3). This inconsistency is largely because doublets are being disproportionately marked/filtered from the lining r2 and r3 populations. The data were reprocessed after reducing the stringency of this step. This results in more consistent (and I think accurate) detection of the lining population, and there is a more reproducible reduction in this population in the CHIKV samples (which supports previous data).
- Plots are included showing expression of the Culemann et al. gene lists used for annotating the macrophage subsets.
- CHIKV+ cells were further assessed. Plots are included for the previous approach where macrophages clusters are split into CHIKV-high/low groups. The enrichment of CHIKV+ cells in specific populations was also assessed using a hypergeometric test. The populations with the highest fraction of CHIKV+ cells are lining and infiltrating macrophages. However, the p-values are not impressive. I’m going to see if I can come up with a better approach for presenting these results.
- Plots are included summarizing differentially expressed genes for mock vs CHIKV for all macrophages grouped together, and for each individual subset.
- There are two key background signals that are observable when
looking at genes downregulated in the CHIKV samples.
- The mock samples are generally lower quality and have a higher proportion of mitochondrial counts, to account for this, mitochondrial genes have been omitted from the figures.
- The mock samples also show higher expression of genes associated with skeletal muscle in all cell types (troponin, myosin genes). In 28 dpi datasets there are two clusters annotated as skeletal muscle, however, these cells are generally too large for the microfluidics used in the 10x platform. Many of these muscle cells are likely not fully intact and the mRNA originating from these cells likely contaminated the buffer during capture. There is no reason to believe this background signal is skewing our results, but this explains why troponin/myosin genes are differentially expressed (for publication it would be reasonable to omit these genes and explain why).
- Plots are included summarizing expression of MHC class II genes for mock vs CHIKV for each subset.
- Plots were regenerated for the Simmons et al. gene lists after adjusting QC filtering as described above.
T cells
- Plots are included for marker genes for different T cell subsets.
- Plots are included showing Ifng expression for all subsets mock vs CHIKV.
- Plots are included summarizing differentially expressed genes for mock vs CHIKV for all T cells grouped together, and for CD4/CD8 T cells grouped together.
UMAP projections show sample groups (left), cell type annotations (middle), and CHIKV+ cells (right).
The fraction of cells passing QC filtering cutoffs is shown for each sample.
Macrophages
Subset annotations
UMAP projections show macrophage subsets.
Module scores for Culemann et al. gene lists are shown for each macrophage cluster. Clusters are colored based on the assigned subset labels.
Heatmaps show mean expression of Culemann et al. gene lists. Replicates are shown separately for each macrophage subset.
CCR2_ARG1_infiltrating
CCR2_IL1B_infiltrating
CX3CR1_lining
RELMa_interstitial
AQP1_interstitial
CHIKV+ cells
UMAP projections show CHIKV+ cells (left) and CHIKV-high clusters (middle, right) for CHIKV-infected samples.
The fraction of cells that are CHIKV+ is shown for each replicate from each macrophage subset. P values were calculated by pooling replicates together, using a hypergeometric test with Bonferroni correction.
The fraction of CHIKV+ cells passing QC filtering cutoffs is shown for each sample.
Differential expression
Top 50 upregulated genes associated with arthritis are shown below for mock- vs CHIKV-infected macrophages (all subsets grouped together).
Genes were filtered for those where the change in expression was in the same direction (up or down) for all replicates, an adjusted p value < 0.05 and an absolute log2 fold change > 0.25 for at least 2/3 replicates
